With the step-by-step capacity expansion of Internet Protocol (IP) service forwarding, a forwarding algorithm for traffic load sharing is becoming more and more important, and directly determines whether network resources can be automatically allocated and employed.
In conventional technology, two load sharing methods are used, including a method for load sharing per stream and a method for load sharing per packet. However, in the conventional load sharing methods, the sharing method and the proportion of the traffic are determined at an inbound interface. In this case, even when an outbound interface for load sharing is congested, the inbound interface would still send the traffic to the outbound interface according to a fixed sharing proportion, which results in traffic congesting and packet loss at the outbound interface.